Bicyclic Substituted Hydroxyphenylmethanones as Novel Inhibitors of

Abbreviations: 17β-HSD1, 17β-hydroxysteroid dehydrogenase type 1; ... of active estrogens by inhibiting their synthetic pathway moves steadily to th...
0 downloads 0 Views 1MB Size
Download PDF
8176 J. Med. Chem. 2010, 53, 8176–8186 DOI: 10.1021/jm101073q

Bicyclic Substituted Hydroxyphenylmethanones as Novel Inhibitors of 17β-Hydroxysteroid Dehydrogenase Type 1 (17β-HSD1) for the Treatment of Estrogen-Dependent Diseases† Alexander Oster, Stefan Hinsberger, Ruth Werth, Sandrine Marchais-Oberwinkler, Martin Frotscher, and Rolf W. Hartmann* Pharmaceutical and Medicinal Chemistry, Saarland University, and Helmholtz Institute for Pharmaceutical Research Saarland (HIPS), Campus C23, D-66123 Saarbr€ ucken, Germany Received August 18, 2010

Estradiol (E2), the most important estrogen in humans, is involved in the initiation and progression of estrogen-dependent diseases such as breast cancer and endometriosis. Its local production in the target cell is regulated by 17β-hydroxysteroid dehydrogenase type 1 (17β-HSD1), which catalyzes E2-formation by reduction of the weak estrogen estrone (E1). Because the enzyme is expressed in the diseased tissues, inhibition of 17β-HSD1 is considered as a promising therapy for the treatment of estrogen-dependent diseases. For the development of novel inhibitors, a structure- and ligand-based design strategy was applied, resulting in bicyclic substituted hydroxyphenylmethanones. In vitro testing revealed high inhibitory potencies toward human placental 17β-HSD1. Compounds were further evaluated with regard to selectivity (17β-HSD2, estrogen receptors ERR and ERβ), intracellular activity (T47D cells), and metabolic stability. The most promising compounds, 14 and 15, showed IC50 values in the low nanomolar range in the cell-free and cellular assays (8-27 nM), more than 30-fold selectivity toward 17β-HSD2 and no affinity toward the ERs. The data obtained make these inhibitors interesting candidates for further preclinical evaluation.

Introduction Breast cancer and endometriosis are estrogen-dependent diseases, which show a strong global prevalence. Breast cancer is one of the two leading causes of cancer death in women. Although some therapeutic approaches are already available, the five-year survival rate for breast cancer patients only remains around 80% (www.cancer.org). The Endometriosis Foundation of America has pointed out that endometriosis is one of the top three causes of female infertility (www. endofound.org). However, until recently, there has been no oral medication available to cure this disease. The local production of estradiol (E2a) plays a pivotal role for the initiation and development of both diseases.1,2 Thus, the reduction of tissue levels of active estrogens by inhibiting their synthetic pathway moves steadily to the forefront. Two therapeutic approaches have mainly been investigated within the past years. Aromatase inhibitors3-8 (to reduce the transformation of androgens into estrogens) and sulfatase inhibitors9 (to reduce the conversion of estrone sulfate (E1S) into estrone (E1)) intervene in the penultimate step of E2 biosynthesis. † For the sake of clarity, IUPAC nomenclature is not strictly followed except for the experimental part where the correct IUPAC names are given. *To whom correspondence should be addressed. Phone: þ49 681 302 70300. Fax: þ49 681 302 70308. E-mail: [email protected]. Website: http://www.pharmmedchem.de. a Abbreviations: 17β-HSD1, 17β-hydroxysteroid dehydrogenase type 1; 17β-HSD2, 17β-hydroxysteroid dehydrogenase type 2; E1, estrone; E2, 17β-estradiol; E1S, estrone sulfate; ER, estrogen receptor; NADP(H), nicotinamide adenine dinucleotide phosphate; NAD(H), nicotinamide adenine dinucleotide; IBX, 2-iodoxybenzoic acid; SAR, structure-activity relationship; RBA, relative binding affinity; SF, selectivity factor; HPLC, high performance liquid chromatography; CC, column chromatography; TLC, thin layer chromatography; IUPAC, international union of pure and applied chemistry.

pubs.acs.org/jmc

Published on Web 10/26/2010

Both therapeutic approaches show disadvantages by their rather radical reduction of systemic estrogen concentration. A softer therapy seems to be the inhibition of 17β-hydroxysteroid dehydrogenase type 1 (17β-HSD1). This enzyme regulates the ratio of the weak active estrogen E1 and its potent metabolite E2 by catalyzing the reduction of E1 to E2 using NAD(P)H as a cofactor (Chart 1). Interestingly, mRNA of 17β-HSD1, which is also used as a prognostic marker, is often overexpressed in breast cancer tissues10,11 and endometriotic lesions.12 Considering the fact that therapeutic approaches regulating intracellular hormone concentrations have already been used successfully in androgen dependent diseases such as benign prostatic hyperplasia and alopecia,13-16 17β-HSD1 appeared to be a promising target for the local reduction of E2 and accordingly for the treatment of breast cancer and endometriosis. In contrast to the applied therapies, a basal estrogenic activity will be maintained because the E1 level should not be affected. Thus, side effects caused by a total absence of estrogenic activity could be avoided. The tissue specific influence of 17β-HSD1 inhibitors seems especially suitable for the therapy of postmenopausal women, in which the local production (e.g., in breast or endometrium) represents the major source of E2. As it has recently been suggested, 17β-HSD1 is also involved in the metabolism of retinoic acid,17,18 a natural compound which is known to have antitumor activity.19 Thus, inhibition of 17β-HSD1 should have an additional beneficial effect. Efficacy of 17β-HSD1 inhibitors has been shown by several groups in different animal models. Husen et al.20,21 and Day et al.22 used xenograft models with MCF7 breast cancer cells stably expressing recombinant human 17β-HSD1 or the breast cancer cell line T47D, respectively, in order to demonstrate in vivo efficacy of their inhibitors by decreasing tumor growth. In r 2010 American Chemical Society

Article

transgenic mice, it was proven that 17β-HSD1 inhibitors are able to reduce the conversion of E1 into E2.23 With regards to endometriosis, two in vivo models have been established.24,25 Using nude mice implanted with endometrial tissue from human donors, the expression of steroid hormone receptors and steroid converting enzymes, as well as the proliferation of the ectopic tissue, can be analyzed.24 In a second model, endometriosis is induced in marmoset monkeys, either by endometrial reflux or invasively by laparotomy.25 Despite encouraging in vivo data, no 17β-HSD1 inhibitor has entered clinical evaluation yet. Thus, the applied inhibitors do not seem to be appropriate candidates for clinical application. The therapeutic concept requires selectivity of potent inhibitors toward 17β-hydroxysteroid dehydrogenase type 2 (17β-HSD2) because it catalyzes the reverse reaction, the oxidation of E2 into E1. Furthermore, to reduce the risk of intrinsic estrogen effects, affinity toward both estrogen receptors R and β (ERR and ERβ) should be avoided, even if the pathophysiological role of ERβ is not completely clarified. The selective inhibition of 17β-HSD1 was the goal of several groups investigating inhibitors over the past years.18,26-30 Their drug design efforts were facilitated by the existence of several X-ray structures of human 17β-HSD1 and will be further alleviated by the recently published first insights in 17β-HSD1 enzyme kinetics and ligand binding by dynamic motion investigations.31 Although running the higher risk of unwanted side effects, most of these groups focused on the development of compounds bearing a steroidal scaffold. Regarding nonsteroidal inhibitors, aside from phytoestrogens27,32 and thiophenepyrimidinones,33,34 there are two additional compound classes showing inhibitory activity in the low nanomolar range, namely the (hydroxyphenyl)naphthols35-39 and the bis(hydroxyphenyl)heterocycles.40-45 Recently, we also reported on heterocyclic substituted biphenylols,46 which demonstrate notable inhibitory activity. Molecular modeling studies indicated that linear and angulate molecules out of this class unfold their activity via interaction with the catalytic center and a rather lipophilic subpocket located below the catalytic site, respectively.46 In the following, we will report on the design, synthesis, and biological evaluation of a novel class of 17β-HSD1 inhibitors, namely bicyclic substituted hydroxyphenylmethanones.

Journal of Medicinal Chemistry, 2010, Vol. 53, No. 22

8177

crystal structure 1EQU (PDB-ID) indicated that their heterocyclic part interacts with different areas of the active site.46 Designed using E1 as a template, compound A shows the same interaction pattern as the natural substrate. In detail: the metaOH-phenyl ring imitates the steroidal A-ring and the methoxypyridine interacts as hydrogen-bond acceptor with Ser142 and Tyr155, which together with Asn114 and Lys159 constitute the catalytic tetrade. While the OH-phenyl ring of B also mimics the steroidal A-ring, its chlorothiophene part points in the direction of a mainly apolar subpocket consisting of Leu95, Leu96, Asn152, Tyr155 and Phe192. Our strategy focuses on the design of inhibitors, which fit into a three-point-pharmacophore model derived from the different interactions found for A and B (Chart 3). A closer analysis of the amino acids in these regions by computational docking studies provides ideas for substituents as appropriate interacting partners as well as for the structural architecture of the compounds. To mimic the interactions established by the steroidal A-ring of the substrate with the amino acids His221 and Glu282, a hydroxy-substituted phenyl moiety will be one important component of our inhibitors. The second part should interact with the catalytic tetrade. Inhibitors with two H-bond acceptor groups located in this region are described to inhibit 17βHSD1.46 One functional group with two hydrogen bond acceptor properties could eventually be more appropriate to target the corresponding amino acids. A keto-group, as in the natural substrate E1, should be the functionality of choice. Moreover, it disrupts the linearity present in A, thus allowing for binding with a third interaction region, which is constituted by a mainly apolar subpocket located beneath the catalytic center. In this subpocket, Asn152 is the only hydrophilic amino acid. This amino acid was already exploited for active compound design: recently, Mazumdar et al.47 reported on the crystal structure of the steroidal inhibitor E2B (Chart 2), whose meta-amidobenzyl substituent was found to interact with Asn152. Consequently, a second OH-phenyl moiety was Chart 3. Design Concept and Pharmacophore Model

Design Molecular docking studies of the most potent heterocyclic substituted biphenylol inhibitors A and B (Chart 2) using the Chart 1. Interconversion of Estrone (E1) and Estradiol (E2)

Chart 2. Most Potent 17β-HSD1 Inhibitors of Heterocyclic Substituted Biphenylol Compound Class (Compounds A and B) and E2B

8178 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 22

introduced as a further component of the inhibitor structure in order to fit in this apolar subpocket and interact with Asn152 (Chart 4). The design concept resulted in compounds 1-5, which were selected as the starting point for this study. Thiophene was chosen as the connecting group between the hydroxyphenyl and the keto-group in order to fit in the rather lipophilic substrate binding site (known as hydrophobic tunnel) and to fit the distance between the various interacting partners in the protein. After identifying 1 and 4 as most promising, small substituents were introduced into these compounds (6-20). Finally, the substitution pattern of the thiophene was changed (21) and it was replaced by thiazole (22-24). Chemistry The synthesis of the thiophene derivatives 1-20 is depicted in Scheme 1. The first step in the preparation of compounds 1-3 and 6-10 was the coupling of 2-bromothiophene with 3-methoxy- and 4-methoxybenzeneboronic acid, respectively, thus leading to intermediates 1b and 6b. A Friedel-Crafts acylation (method A) was then performed with the appropriate Chart 4. Schematic Representation of the Inhibitor Template Binding to the Amino Acids of the Active Site (Possible Hydrogen Bond Interactions Are Depicted in Dashed Lines)

Oster et al.

benzoyl chlorides, resulting in compounds 1a-3a and 6a-10a. Subsequently, treatment with boron tribromide led to the corresponding dihydroxylated final compounds 1-3 and 6-10. The brominated key intermediate 4b was synthesized via Friedel-Crafts acylation of 2-bromothiophene with 3-methoxybenzoyl chloride. Using 4b, Suzuki cross coupling reactions48 with the appropriate commercially available boronic acids were carried out via two different methods (B for compounds 14a-16a and C41 for compounds 4a, 5a, 11a-13a, and 18a, Scheme 1). The methoxy functions were cleaved with BBr349 yielding compounds 4, 5, 11-16, and 18. Ether cleavage of the key intermediate 4b gave 17a in quantitative yield. Cross coupling of 17a with the appropriate hydroxylated benzene boronic acids resulted in the dihydroxylated compounds 17 and 19. 3-Ethyl-4-methoxybenzeneboronic acid was synthesized using the method of Allan et al.50 and was used for Suzuki reaction in order to yield the monomethoxy derivative 20a. The latter was heated with pyridinium hydrochloride51 to yield compound 20. The preparation of compounds 21-24 (Scheme 2) started with a Grignard reaction using 3-methoxyphenylmagnesium bromide and the appropriate brominated aromatic carbaldehydes (method F). The OH-groups of compounds 21c-23c were subsequently oxidized to the keto-functions with 2iodoxybenzoic acid (IBX) (method G). Suzuki cross coupling reactions using the 4-methoxy-3-methylbenzeneboronic acid and the brominated methanones 21b-23b led to compounds 21a-23a. Demethylation with BBr3 (method D) yielded the final compounds 21 and 23, whereas pyridinium hydrochloride (method E) was used for the preparation of 22. For the preparation of 24, Suzuki coupling was performed using the hydroxylated intermediate 24c because its oxidation failed. Subsequently, 24b was oxidized by means of IBX in anhydrous THF. Reaction of 24a with pyridinium hydrochloride led to 24. Biological Results Activity: Inhibition of Human 17β-HSD1. Placental enzyme was isolated following a described procedure38 and

Scheme 1. Synthesis of Compounds 1-20a

Reagents and conditions: (a) method A, AlCl3, anhydrous CH2Cl2, 0 °C, 0.5 h and then rt, 1 h, for compounds 1a-3a, 4b, 6a-10a; (b) method B, Cs2CO3, Pd(PPh3)4, DME/water (1:1), reflux, 2 h, for compounds 14a-16a, 17, 19, 20a; (c) method C, Cs2CO3, Pd(PPh3)4, DME/EtOH/water (1:1:1), microwave conditions (150 W, 15 bar, 150 °C, 15 min), for compounds 4a, 5a, 11a-13a, 18a; (d) method D, BBr3, CH2Cl2, -78 °C, 20 h, for compounds 1-16, 17a, 18; (e) method E, pyridinium hydrochloride, 220 °C, 18 h, for compound 20; (f) Na2CO3, Pd(PPh3)4, THF/water (1:1), reflux, 18 h, for compound 1b. a

Article

Journal of Medicinal Chemistry, 2010, Vol. 53, No. 22

8179

Scheme 2. Synthesis of Compounds 21-24a

a Reagents and conditions: (a) method F, anhydrous THF, 80 °C, 3 h, for compounds 21c-24c; (b) method G, 2-iodoxybenzoic acid, anhydrous THF, 0 °C, 10 min and 60 °C, 18 h, for compounds 21b-23b, 24a; (c) method B, Cs2CO3, Pd(PPh3)4, DME/water (1:1), reflux, 2 h, for compounds 21a-23a; (d) method D, BBr3, CH2Cl2, -78 °C, 20 h, for compounds 21, 23; (e) method E, pyridinium hydrochloride, 220 °C, 18 h, for compounds 22, 24; (f) Cs2CO3, Pd(PPh3)4, DME/water (1:1), reflux, 4 h, for compound 24b

incubated with tritiated E1, cofactor, and inhibitor. After HPLC separation of substrate and product, the amount of labeled E2 formed was quantified. The inhibition values of compounds 1-24 are presented in Tables 1 and 2. Compounds 1-5 with their keto and two OH functions were designed to interact with three distinct areas of the active site. Among these inhibitors, the meta-meta substituted 1 showed the best IC50 value (22 nM), demonstrating that in this class of compounds, the meta-OH substituted phenyl moiety is also very relevant for 17β-HSD1 inhibition as observed with the bis(hydroxyphenyl)heterocycles.41 Displacement of the hydroxy group on the benzoyl ring led to a decrease in activity (2 and 3). The position of the OH group at the phenyl ring has a smaller impact on inhibitory activity: para-OH 4 (IC50 = 33 nM) was equipotent to meta-OH 1, the potency of ortho-OH 5 was only slightly reduced (IC50 =95 nM). The best substitution patterns identified in 1 (meta-meta) and 4 (meta-para) were sustained in the design of the further compounds. To examine the influence of additional small substituents in paraposition of the benzoyl group, compounds 6-10 were synthesized. Introduction of a methyl substituent decreased the inhibitory activity in both series strongly (6 and 8). An exchange of methyl by fluorine did not affect inhibition (7 and 9). Introduction of methyl in para-position of the phenyl ring led to a 10-fold decrease in activity (11), whereas meta-CH3, para-OH 14 turned out to be a very potent compound (IC50 = 8 nM). Fluorinated compounds with different substitution patterns showed a similarly high inhibition of 17β-HSD1 (IC50s: 12, 30 nM; 13, 6 nM; 15, 19 nM). It is remarkable that the substitution with fluorine, in contrast to methyl, does not decrease the high inhibitory potency of the parent compounds 1 and 4, regardless of their position at the benzoyl or phenyl ring. To provide more insight into the role of different substituents, 14 and 15 were used as starting points for further structural modifications. To combine the characteristics of methyl and fluorine, a CF3 group (16), as well as chlorine (17) as its bioisosteric equivalent, was introduced. Both inhibitors confirmed the expected high inhibitory activity with IC50 values of 14 and 5 nM, respectively. In contrast to this finding, a slight decrease in activity was caused by the

replacement of the CH3 for a hydroxy (18: IC50 = 86 nM) or a methoxy (19: IC50 = 108 nM). Interestingly, the ethyl derivative 20 once again revealed a high biological potency (IC50 = 20 nM), thus indicating that an oxygen in this position impinges on enzyme affinity. Originally chosen as a suitable lipophilic component to fit the mainly apolar substrate binding site, we replaced the 2,5-disubstituted thiophene with 2,4-disubstituted thiophene and thiazole (21-24) in order to investigate the relevance of the middle ring on inhibitory activity. In the structurally related class of bis(hydroxyphenyl)heterocycles, a strong correlation between the nature of the heterocycle and the affinity toward 17β-HSD1 was already shown.41 As 14 is one of the most active inhibitors in this series, its substitution pattern was maintained for this study. The biological results of compounds 21-24 are shown in Table 2. With regard to 2,4-disubstituted 21, the decrease of activity (IC50 =199 nM) underlines the importance of the sulfur position. Exchanging thiophene with thiazole led to compounds 22-24. Compound 22, which bears the sulfur of the thiazole in the same position as 14, the thiophene-sulfur, still demonstrated a very good activity (IC50 =35 nM). Interestingly, the potency of 23 and 24 strongly decreased, thereby confirming that the position of the sulfur plays a pivotal role for 17β-HSD1 inhibition. Selectivity: Inhibition of Human 17β-HSD2 and Affinities to ERr and ERβ. Inhibition of 17β-HSD type 2 must be avoided, as it acts as a biological counterpart by catalyzing the reverse reaction: oxidation of E2 to E1. Compared with the type 1 enzyme, inhibition of 17β-HSD2 was determined in a similar assay using placental microsomes. These were incubated with tritiated E2 in the presence of NADþ and the inhibitor. The separation and quantification of the labeled product was performed by HPLC using radiodetection. IC50 values and the percentage of inhibition, respectively, are shown in Tables 1 and 2. In vitro selectivity toward 17β-HSD2 is also expressed as selectivity factor (SF) describing the ratio of the concentration required to inhibit the activity of the isoenzymes by 50% (IC50 17β-HSD2/IC50 17β-HSD1).

8180 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 22

Oster et al.

Table 1. Inhibition of Human 17β-HSD1 and 17β-HSD2 by Compounds 1-20

IC50 [nM]a compd

R1

1 2 3 4 5

3-OH 4-OH 2-OH 3-OH 3-OH

1 8 9 10 11 12 13 4 6 7 14 15 16 17 18 19 20

R2 3-OH 3-OH 3-OH 4-OH 2-OH

17β-HSD1b

17β-HSD2c

22 368 945 33 95

109 376 567 478 18

4-CH3 4-F 2-F

22 64%e 18 594 207 30 6

109 31%e 49 240 465 57 19

CH3 F CF3 Cl OH OCH3 CH2CH3

33 52%e 21 8 19 16 5 86 108 20

478 13%e 69 382 588 95 246 590 793 786

CH3 F NO2

CH3 F

SFd 5 1 0.6 14 0.2 5 3 0.4 2 2 3 14 3 48 31 6 49 7 7 39

a Mean value of three determinations, standard deviation less than 23%. b Human placenta, cytosolic fraction, substrate E1, 500 nM, cofactor NADH, 500 μM. c Human placenta, microsomal fraction, substrate E2, 500 nM, cofactor NADþ, 1500 μM. d Selectivity factor: IC50 (17β-HSD2)/IC50 (17β-HSD1). e Inhibition at 1 μM (inhibitor concentration).

Table 2. Inhibition of Human 17β-HSD1 and 17β-HSD2 by Compounds 14, 21-24

a

Mean value of three determinations, standard deviation less than 14%. b Human placenta, cytosolic fraction, substrate E1, 500 nM, cofactor NADH, 500 μM. c Human placenta, microsomal fraction, substrate E2, 500 nM, cofactor NADþ, 1500 μM. d selectivity factor: IC50 (17β-HSD2)/IC50 (17β-HSD1). e Inhibition at 1 μM (inhibitor concentration).

Regarding compounds 1-5, it became apparent that just the modification of the OH-substitution pattern influences the selectivity toward the type 2 enzyme. While compound 4 showed the best SF (14) out of these five, compounds 3 and 5, which bear one of their OH functions in the ortho position,

turned out to be potent inhibitors of 17β-HSD2 (SF = 0.6 and SF=0.2, respectively). Introduction of further substituents on the benzoyl ring did not strongly influence the SF, with the exception of compound 10, whose nitro group is also responsible for the reverse selectivity in favor of 17β-HSD2 (SF=0.4). Concerning the different phenyl ring modified compounds, insertion of CH3 (14, SF=48) and F (15, SF =31) in the meta position (ortho to OH) increased selectivity compared to the unsubstituted compound 4 (SF = 14). While other positions had less of a high impact (11-13), this meta position seems to be a key position for 17β-HSD1 activity and selectivity toward 17β-HSD2. Chlorine (17) and ethyl (20) showed comparable high selectivities (SFs=49 for 17 and 39 for 20). The fact that chlorine and trifluoromethyl substituents have comparable electronic influences on a phenyl ring implies that other parameters are responsible for the difference in selectivity between 17β-HSD1 and 2. Regarding the influence of the different middle rings on selectivity toward 17β-HSD2, it is striking that when compared with 14, the isomeric thiophene 21 showed a strong reduction of selectivity. Only thiazole 22 showed a comparatively high SF of 14. Surprisingly, thiazoles 23 and 24 turned out to be more suitable inhibitors for 17β-HSD2, thereby demonstrating that the nature of the middle ring has a major impact on activity and selectivity. To keep the risk of unwanted side effects as low as possible, intrinsic estrogenic effects should be avoided. Therefore, it is a

Article

Journal of Medicinal Chemistry, 2010, Vol. 53, No. 22

8181

Table 3. Inhibition of 17β-HSD1 in Cell-Free and Cellular Assay by Selected Compounds

a Mean values of three determinations, standard deviation less than 23%. b Human placental, cytosolic fraction, substrate E1, 500 nM, cofactor NADH, 500 μM. c T47D cells, substrate E1, 50 nM.

prerequisite for 17β-HSD1 inhibitors to show no or low affinity for the ERs. Relative binding affinities (RBA) of compounds, which showed a selectivity factor higher than 10, were determined using recombinant human protein in a competitive binding assay applying tritium labeled E2 and hydroxyapatite. Because RBA of E2 is equated with 100%, compounds showing a RBA value below 0.1% are considered to have neglectable receptor affinity. Within this study, none of the evaluated inhibitors crossed this value, neither for ERR nor ERβ. Further Biological Evaluations. Besides the cell-free assay, T47D, a breast cancer cell line reflecting the conditions in many breast tumors by expressing subtypes 1 and 2 of 17β-HSD, was used to determine the intracellular potency of compounds having a SF higher than 10 (Table 3). While inhibitors 4, 14, and 15 confirmed their strong activity in T47D-cells with IC50 values in the low nanomolar range (between 17 and 30 nM), compounds 17 and 22 showed inhibitory activities in the 100 nM range. Compared to the cell-free assay, only chloro-(17) and ethylderivative (20) demonstrated a remarkable reduction of inhibitory activity in T47D cells. Additionally, metabolic stability of compounds 14, 17, and 22 was evaluated using human liver microsomes. The corresponding data of the selected compounds and the reference compounds dextromethorphan and verapamil are presented in Table 4. All three inhibitors can be classified in the medium clearance category according to their half-life times around 100 min and thus are considered to have a reasonable metabolic stability. None of the small variations among these three inhibitors caused considerable differences in liability of the compounds to metabolic degradation.

Discussion and Conclusions The results of molecular modeling studies in the class of heterocyclic substituted biphenylols46 provided the basis for a successful design strategy using a new three-point-pharmacophore model. The determined biological results place emphasis on the validity of this structure- and ligand-based design approach, resulting in compound 1 as a new lead for further optimization. Interestingly, the position of a hydroxy function at the benzoyl ring has more influence on inhibitory activity than the one at the phenyl ring. This suggests that either the expected hydrogen-bond interactions of the latter with His221 and Glu282 are less important for 17β-HSD1 inhibition or that the protein is flexible enough, at least in this part of the enzyme, to allow equal interaction with each derivative 1, 4, and 5. In contrast to previous studies,43 an exchange of this OH-phenyl moiety might also lead to highly active 17β-HSD1 inhibitors. The introduction of further substituents, such as methyl and fluorine in different positions on the benzoyl and phenyl ring, resulted in interesting modulations of biological activities. Compounds 6 and 8, bearing the methyl substituent in the para position of the benzoyl ring, showed a strong decrease of inhibitory activity, once again emphasizing the sharp structureactivity relationships within the rather apolar subpocket. There is plausibly a sterical hindrance caused by the CH3 group, thus leading to a shifted position of the molecule that is unable to establish suitable hydrogen bond interactions. Taking into account that the geometrical radius of fluorine is smaller than methyl, the fluorine substituted hydroxyphenyl-part of

8182 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 22 Table 4. Metabolic Stability of Compounds 14, 17, and 22 and Reference Compounds Using Human Liver Microsomes compd

CLint [μL/min/mg protein]

t1/2 [min]

dextrometorphana verapamila 14a 17b 22a

10.7 105 13.8 12.9 17.0

129 13.2 100 108 81.4

a

Test concentration = 3 μM. b Test concentration = 0.5 μM.

compounds 7 and 9 obviously still fits well into the subpocket, as can be seen by their high inhibitory activities. Regarding the hydroxyphenyl, the introduction of methyl and fluorine in ortho- to the para-OH substituent (14 and 15) improved inhibitory activity and even selectivity toward 17βHSD2. It is striking that derivatives with more polar substituents (OH in compound 18 and OCH3 in compound 19) in this position could not exert such a high inhibitory activity, and especially their selectivity factor decreased. The fact that the Cl and CH2CH3 compounds 17 and 20 have similar inhibitory properties comparable to those of 14 and 15, led to the assumption that the amino acids surrounding this region of the 17β-HSD1 binding site are predominantly lipophilic. Only compound 16, which bears a CF3 group, demonstrated a low selectivity factor because of its strong inhibition of 17β-HSD2. The ethyl derivative 20, which has a marginally lower SF than 14, revealed a slightly reduced 17β-HSD2 inhibition, indicating that an enlargement of this lipophilic substituent might lead to lower inhibition of the type 2 enzyme. Concerning selectivity toward ERR and ERβ, each inhibitor out of this new compound class showed neglectable affinity to the ERs, which confirms our strategy of developing nonsteroidal inhibitors. Despite the fact that the keto function and one hydroxy group of the bicyclic substituted hydroxyphenylmethanones have to be considered as mimicks of the oxygen functions of E1, obviously, the absence of the sterane core and the additional phenolic ring are the main responsible factors that we succeeded in obtaining compounds without ER affinity. While the presence of the keto linker is the only structural difference between this new class of 17β-HSD1 inhibitors and the previously described bis(hydroxyphenyl)heterocycles,40-43 several structure-activity relationships are in accordance. For example, the influence of the (hetero)aromatic middle ring is comparable. The exchange of the 2,5-disubstituted thiophene with 2,4-disubstituted thiophene 21 or different thiazoles 22-24 revealed that the position of the sulfur plays a critical role for enzyme inhibition. However, only compounds in which the sulfur is located between the two substituents show inhibition values in the low nanomolar range. Because it was shown that 2,4-thiophenes are also able to exhibit strong 17βHSD1 inhibition,41 the molecular electrostatic potential of the inhibitors can be discussed as a possible reason for the differences in inhibitory activity.41 With regards to thiazoles, the corresponding compounds in the class of bis(hydroxyphenyl)heterocycles also revealed similar results. In both studies, there is a rather radical loss of inhibitory activity for all thiazole derivatives with nitrogen located between the two substituents, thus pointing out that this substitution pattern of thiazoles is detrimental for 17β-HSD1 inhibition. Interestingly, while compound 22 resembles the best inhibitor 14 in activity and selectivity, compounds 23 and 24 inhibited significantly stronger 17β-HSD2 than the type 1 enzyme. As there is neither an X-ray structure nor an applicable homology model

Oster et al.

of 17β-HSD2 available, these results cannot be interpreted on a sound basis. The most important advantage of this compound class is its high cellular inhibitory activity. While most of the described intracellularly active 17β-HSD1 inhibitors were tested using low substrate concentrations (2-30 nM),33,50 the new compound class revealed IC50 values in the low nanomolar range in spite of using a much higher substrate concentration of 50 nM. Evaluated in our assay, the most potent inhibitors of the bis(hydroxyphenyl)heterocycles showed cellular inhibitory activities around 300 nM (IC50).43 Therefore, the keto function seems to have a positive influence on cell permeation, protein binding, and/or intracellular metabolism. Only compounds 17 and 20 showed intracellularly a reduction of inhibitory activity compared to the cell-free assay. Increased lipophilicity might be responsible for this finding. Concerning the metabolic stability study, the tested inhibitors 14, 17, and 22 revealed a medium intrinsic clearance confirming a reasonable metabolic profile. The fact that all three inhibitors showed comparable profiles indicates that neither the exchange thiophene-thiazole nor a variation of the additional substituent next to the hydroxy function can give rise to some significant metabolic liability. In this paper, we described the structure-based design approach, synthesis, and biological evaluations of bicyclic substituted hydroxyphenylmethanones as new highly potent and selective inhibitors of 17β-HSD1. The applied design concept is based on the newly built pharmacophore model, which resulted in compound 1 as the starting point of the study. The position of the hydroxy functions, the influence of further small substituents and an alternative central ring were investigated. Structural optimization of this lead led to the development of 14, 15, 17, and 22, all of which show inhibitory activities in the very low nanomolar range in cell-free and cellular assays, remarkable selectivity toward 17β-HSD2, no affinity on both ERs, and very good metabolic stability (proven in case of 14, 17, and 22) tested on human liver microsomes. Hence, they represent promising 17β-HSD1 inhibitors, which fulfill all prerequisites for consideration as clinical candidates. Nevertheless, further optimizations appear possible, especially with regards to the phenyl ring and will therefore be investigated in a future study. Experimental Section Chemical Methods. Chemical names follow IUPAC nomenclature. Starting materials were purchased from Aldrich, Acros, Lancaster, Maybridge, Combi Blocks, Merck, or Fluka and were used without purification. Column chromatography (CC) was performed on silica gel (70-200 μm), preparative thin layer chromatography (TLC) on 1 mm SIL G-100 UV254 glass plates (Macherey-Nagel), and reaction progress was monitored by TLC on Alugram SIL G UV254 (Macherey-Nagel). All microwave irradiation experiments were carried out in a CEM-Discover monomode microwave apparatus. 1 H NMR and 13C NMR spectra were measured on a Bruker AM500 spectrometer (500 MHz) at 300 K. Chemical shifts are reported in δ (parts per million, ppm), by reference to the hydrogenated residues of deuteriated solvent as internal standard (CDCl3, δ=7.24 ppm (1H NMR) and δ=77 ppm (13C NMR); CD3OD, δ = 5.84 ppm (1H NMR) and δ = 49.3 ppm (13C NMR); CD3COCD3, δ = 2.05 ppm (1H NMR) and δ = 30.8 ppm (13C NMR); CD3SOCD3, δ = 2.50 ppm (1H NMR) and δ = 39.5 ppm (13C NMR)). Signals are described as s, d, t, dd, ddd, m, dt, td, and q for singlet, doublet, triplet, doublet of doublets, doublet

Article

of doublets of doublets, multiplet, doublet of triplets, triplet of doublets, and quadruplet, respectively. All coupling constants (J) are given in hertz (Hz). Mass spectra (ESI) were recorded on a TSQ Quantum (Thermo Finnigan) instrument. Tested compounds have >95% chemical purity as measured by HPLC. The methods for HPLC analysis and a table of data for all tested compounds are provided in the Supporting Information. The following compounds were prepared according to previously described procedures: 4-fluoro-3-methoxybenzoyl chloride,52 4-nitro-3-methoxybenzoyl chloride,52 3-ethyl-4-methoxybenzeneboronic acid (20b),50 2-(4-methoxyphenyl)thiophene (6b).53 General Procedure for Friedel-Crafts Acylation. Method A. A mixture of monosubstituted thiophene derivative (1 equiv), arylcarbonyl chloride (1 equiv), and aluminumtrichloride (1 equiv) in anhydrous dichloromethane was stirred at 0 °C for 0.5 h. The reaction mixture was warmed to room temperature and stirred for 1 h. 1 M HCl was used to quench the reaction. The aqueous layer was extracted with ethyl acetate. The combined organic layers were washed with brine, dried over magnesium sulfate, filtered, and concentrated to dryness. The product was purified by CC. General Procedures for Suzuki Coupling. Method B. A mixture of arylbromide (1 equiv), boronic acid derivative (1.2 equiv), cesium carbonate (4 equiv), and tetrakis(triphenylphosphine) palladium (0.01 equiv) in an oxygen free DME/water (1:1) solution was refluxed under nitrogen atmosphere for 2 h. The reaction mixture was cooled to room temperature. The aqueous layer was extracted with ethyl acetate. The combined organic layers were washed with brine, dried over magnesium sulfate, filtered, and concentrated to dryness. The product was purified by CC (except for 17 (CC followed by preparative HPLC) and 19 (CC followed by preparative TLC)). Method C. A mixture of arylbromide (1 equiv), boronic acid derivative (1.2 equiv), cesium carbonate (4 equiv), and tetrakis(triphenylphosphine) palladium (0.01 equiv) was suspended in an oxygen free DME/EtOH/water (1:1:1) solution. The reaction mixture was exposed to microwave irradiation (15 min, 150 W, 150 °C, 15 bar). After reaching room temperature, water was added and the aqueous layer was extracted with ethyl acetate. The combined organic layers were washed with brine, dried over magnesium sulfate, filtered, and concentrated to dryness. The product was purified by CC. General Procedures for Ether Cleavage. Method D. To a solution of methoxybenzene derivative (1 equiv) in anhydrous dichloromethane at -78 °C (dry ice/acetone bath), boron tribromide in dichloromethane (1 M, 3 eq per methoxy function) was added dropwise. The reaction mixture was stirred for 20 h at room temperature under nitrogen atmosphere. Water was added to quench the reaction, and the aqueous layer was extracted with ethyl acetate. The combined organic layers were washed with brine, dried over sodium sulfate, filtered, and concentrated to dryness. The product was purified by CC, CC followed by preparative TLC, or CC followed by preparative HPLC, respectively. Method E. A mixture of methoxybenzene derivative (1 equiv) and pyridinium hydrochloride (37 equiv per methoxy function) was heated to 220 °C for 18 h. After cooling to room temperature, water, 1 M HCl and ethyl acetate were added. The aqueous layer was extracted with ethyl acetate. The combined organic layers were washed with brine, dried over sodium sulfate, filtered, and concentrated to dryness. The product was purified by CC followed by preparative TLC or preparative HPLC, respectively. General Procedure for Grignard Reaction. Method F. To a solution of carbaldehyde derivative (1 equiv) in anhydrous THF, 3-methoxyphenylmagnesium bromide (1.0 M in THF/toluene, 2.2 equiv) was added dropwise. The reaction mixture was stirred for 3 h at 80 °C under nitrogen atmosphere. Brine was added to quench the reaction, and the aqueous layer was extracted with ethyl acetate. The combined organic layers were washed with

Journal of Medicinal Chemistry, 2010, Vol. 53, No. 22

8183

brine, dried over magnesium sulfate, filtered, and concentrated to dryness. The product was purified by CC. General Procedure for Oxidation. Method G. A mixture of aliphatic alcohol derivative (1 equiv) and 2-iodoxybenzoic acid (2 equiv) in anhydrous THF was stirred at 0 °C. After 10 min, the reaction mixture was stirred and heated to 60 °C for 18 h. After cooling to room temperature, saturated sodium thiosulfate solution was added to quench the reaction and the aqueous layer was extracted with ethyl acetate. The combined organic layers were washed succesively with 1 M NaOH and brine, dried over magnesium sulfate, filtered, and concentrated to dryness. The product was purified by CC. General Procedure for Purification Using Preparative HPLC. All declared compounds were purified via an Agilent Technologies series 1200-preparative HPLC using a RP C18 Nucleodur 100-5 column (30 mm  100 mm/50 μM from Macherey Nagel GmbH) as stationary phase with a linear gradient run (solvents: acetonitrile, water) starting from 20% acetonitrile up to 100% in 36 min. (3-Hydroxyphenyl)[5-(3-hydroxyphenyl)-2-thienyl]methanone (1). The title compound was prepared by reaction of (3-methoxyphenyl)[5-(3-methoxyphenyl)-2-thienyl]methanone (1a) (138 mg, 0.43 mmol) and boron tribromide (2.58 mmol) according to method D. The product was purified by CC (hexane/ethyl acetate 8:2) followed by preparative HPLC; yield: 75% (95 mg). 1H NMR (CD3OD): 7.64 (d, J = 4.1 Hz, 1H), 7.42 (d, J = 4.1 Hz, 1H), 7.33 (t, J = 7.9 Hz, 1H), 7.27 (dt, J = 1.3 Hz and J = 7.6 Hz, 1H), 7.24-7.17 (m, 3H), 7.12 (t, J = 2.1 Hz, 1H), 7.03 (ddd, J = 0.9 Hz and J = 2.5 Hz and J = 7.9 Hz, 1H), 6.80 (ddd, J = 0.9 Hz and J = 2.5 Hz and J = 7.9 Hz, 1H). 13C NMR (CD3OD): 189.90, 159.40, 159.00, 155.05, 140.60, 139.10, 137.90, 131.45, 130.80, 125.40, 121.30, 120.70, 118.60, 117.50, 116.55, 114.00. (3-Hydroxyphenyl)[5-(4-hydroxyphenyl)-2-thienyl]methanone (4). The title compound was prepared by reaction of (3-methoxyphenyl)[5-(4-methoxyphenyl)-2-thienyl]methanone (4a) (178 mg, 0.55 mmol) and boron tribromide (3.30 mmol) according to method D. The product was purified by CC (hexane/ethyl acetate 8:2) followed by preparative TLC (hexane/ethyl acetate 1:1); yield: 51% (83 mg). 1H NMR (CD3COCD3): 8.79 (s, 1H), 8.72 (s, 1H), 7.67-7.64 (m, 3H), 7.42 (d, J = 4.1 Hz, 1H), 7.37 (t, J = 7.9 Hz, 1H), 7.35 (td, J = 1.3 Hz and J = 7.6 Hz, 1H), 7.33-7.32 (m, 1H), 7.11 (ddd, J = 0.9 Hz and J = 2.5 Hz and J = 7.9 Hz, 1H), 6.95 (d, J = 8.8 Hz, 1H). 13C NMR (CD3COCD3): 188.60, 160.60, 159.35, 155.20, 142.80, 141.50, 138.05, 131.50, 129.60, 126.80, 124.70, 122.00, 121.05, 117.95, 117.25. [5-(4-Hydroxy-3-methylphenyl)-2-thienyl](3-hydroxyphenyl)methanone (14). The title compound was prepared by reaction of [5-(4-methoxy-3-methylphenyl)-2-thienyl](3-methoxyphenyl)methanone (14a) (180 mg, 0.53 mmol) and boron tribromide (3.18 mmol) according to method D. The product was purified by CC (hexane/ethyl acetate 7:3); yield: 90% (148 mg). 1H NMR (CD3COCD3): 8.68 (s, 1H), 8.66 (s, 1H), 7.65 (d, J = 4.1 Hz, 1H), 7.57-7.56 (m, 1H), 7.47 (dd, J = 2.2 Hz and J = 8.2 Hz, 1H), 7.41 (d, J = 4.1 Hz, 1H), 7.39 (t, J = 7.8 Hz, 1H), 7.35 (dt, J = 1.3 Hz and J = 7.6 Hz, 1H), 7.32-7.31 (m, 1H), 7.11 (ddd, J = 1.3 Hz and J = 2.5 Hz and J = 7.9 Hz, 1H), 6.92 (d, J = 8.2 Hz, 1H), 2.27 (s, 3H). 13C NMR (CD3COCD3): 188.55, 159.35, 158.65, 155.50, 142.65, 141.60, 138.05, 131.50, 130.70, 127.25, 126.90, 126.75, 124.55, 122.00, 121.00, 117.25, 117.20, 17.10. [5-(3-Fluoro-4-hydroxyphenyl)-2-thienyl](3-hydroxyphenyl)methanone (15). The title compound was prepared by reaction of [5-(3-fluoro-4-methoxyphenyl)-2-thienyl](3-methoxyphenyl)methanone (15a) (200 mg, 0.58 mmol) and boron tribromide (3.48 mmol) according to method D. The product was purified by CC (hexane/ethyl acetate 7:3) followed by preparative TLC (hexane/ ethyl acetate 6:4); yield: 90% (164 mg). 1H NMR (CD3COCD3): 9.09 (s, 1H), 8.74 (s, 1H), 7.66 (d, J = 4.1 Hz, 1H), 7.56 (dd, J = 2.2 Hz and J = 12.3 Hz, 1H), 7.47-7.44 (m, 2H), 7.39 (t, J = 7.8 Hz, 1H), 7.35 (dt, J = 1.5 Hz and J = 7.6 Hz, 1H), 7.34-7.32

8184 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 22

(m, 1H), 7.13 (ddd, J = 1.3 Hz and J = 2.5 Hz and J = 7.8 Hz, 1H), 7.09 (t, J = 8.8 Hz, 1H). 13C NMR (CD3COCD3): 188.65, 159.35, 154.45, 153.45, 147.80, 143.50, 141.30, 137.95, 131.55, 127.50, 125.65, 124.70, 122.05, 121.15, 120.35, 117.25, 115.60. [5-(3-Chloro-4-hydroxyphenyl)-2-thienyl](3-hydroxyphenyl)methanone (17). The title compound was prepared by reaction of (5-bromo-2-thienyl)(3-hydroxyphenyl)methanone (17a) (150 mg, 0.53 mmol), 3-chloro-4-hydroxybenzene boronic acid (110 mg, 0.64 mmol), cesium carbonate (691 mg, 2.12 mmol), and tetrakis(triphenylphosphine) palladium (6 mg, 5 μmol) according to method B. The product was purified by CC (hexane/ethyl acetate 8:2) followed by preparative HPLC; yield: 42% (73 mg). 1H NMR (CD3COCD3): 7.79 (d, J = 2.2 Hz, 1H), 7.68 (d, J = 4.1 Hz, 1H), 7.60 (dd, J = 2.2 Hz and J = 8.5 Hz, 1H), 7.51 (d, J = 4.1 Hz, 1H), 7.39 (td, J = 0.6 Hz and J = 7.8 Hz, 1H), 7.35 (dt, J = 1.5 Hz and J = 7.8 Hz, 1H), 7.33-7.32 (m, 1H), 7.13-7.11 (m, 2H). 13C NMR (CD3COCD3): 188.60, 159.40, 156.05, 153.20, 143.55, 141.35, 137.95, 131.55, 129.40, 128.10, 128.00, 125.65, 123.10, 122.00, 121.15, 119.30, 117.25. [5-(3-Ethyl-4-hydroxyphenyl)-2-thienyl](3-hydroxyphenyl)methanone (20). The title compound was prepared by reaction of [5-(3-ethyl-4-methoxyphenyl)-2-thienyl](3-hydroxyphenyl)methanone (20a) (116 mg, 0.34 mmol) and pyridinium hydrochloride (1.45 g, 12.58 mmol) according to method E. The product was purified by CC (hexane/ethyl acetate 6:4) followed by preparative TLC (hexane/ethyl acetate 6:4); yield: 47% (52 mg). 1H NMR (CD3COCD3): 8.72 (s, 1H), 8.68 (s, 1H), 7.66 (d, J=4.1 Hz, 1H), 7.57 (d, J =2.5 Hz, 1H), 7.47 (d, J =2.5 Hz and J= 8.2 Hz, 1H), 7.42 (d, J =4.1 Hz, 1H), 7.39 (t, J =7.9 Hz, 1H), 7.34 (dt, J =1.5 Hz and J =7.6 Hz, 1H), 7.32-7.31 (m, 1H), 7.11 (ddd, J =1.3 Hz and J =2.5 Hz and J =7.9 Hz, 1H), 6.93 (d, J =8.2 Hz, 1H), 2.70 (q, J =7.6 Hz, 2H), 1.24 (t, J =7.6 Hz, 3H). 13C NMR (CD3COCD3): 187.60, 158.35, 157.30, 154.65, 149.55, 141.65, 140.60, 137.05, 132.35, 130.50, 128.25, 125.90, 123.60, 121.00, 120.00, 116.50, 116.25, 23.90, 14.50. [2-(4-Hydroxy-3-methylphenyl)-1,3-thiazol-5-yl](3-hydroxyphenyl)methanone (22). The title compound was prepared by reaction of [2-(4-methoxy-3-methylphenyl)-1,3-thiazol-5-yl](3-methoxyphenyl)methanone (22a) (65 mg, 0.19 mmol) and pyridinium hydrochloride (1.63 g, 14.06 mmol) according to method E. The product was purified by CC (hexane/ethyl acetate 8:2) followed by preparative TLC (hexane/ethyl acetate 6:4); yield: 46% (27 mg). 1 H NMR (CD3COCD3): 8.26 (s, 1H), 7.87-7.86 (m, 1H), 7.78 (dd, J = 2.2 Hz and J = 8.5 Hz, 1H), 7.43-7.40 (m, 2H), 7.387.37 (m, 1H), 7.16-7.14 (m, 1H), 6.97 (d, J = 8.2 Hz, 1H), 2.29 (s, 3H). 13C NMR (CD3COCD3): 188.25, 176.30, 161.20, 159.65, 151.65, 141.25, 131.75, 131.60, 128.20, 127.40, 126.35, 121.90, 121.65, 117.20, 17.15. Biological Methods. [2,4,6,7-3H]-E2 and [2,4,6,7-3H]-E1 were bought from Perkin-Elmer, Boston. Quickszint Flow 302 scintillator fluid was bought from Zinsser Analytic, Frankfurt. 17β-HSD1 and 17β-HSD2 were obtained from human placenta according to previously described procedures.38 Fresh human placenta was homogenized, and cytosolic fraction and microsomes were separated by centrifugation. For the partial purification of 17β-HSD1, the cytosolic fraction was precipitated with ammonium sulfate. 17β-HSD2 was obtained from the microsomal fraction. 1. Inhibition of 17β-HSD1. Inhibitory activities were evaluated by an established method with minor modifications.38 Briefly, the enzyme preparation was incubated with NADH (500 μM) in the presence of potential inhibitors at 37 °C in a phosphate buffer pH 7.4 (50 mM) supplemented with 20% of glycerol and EDTA (1 mM). Inhibitor stock solutions were prepared in DMSO. The final concentration of DMSO was adjusted to 1% in all samples. The enzymatic reaction was started by addition of a mixture of unlabeled- and [2,4,6,7-3H]-E1 (final concentration: 500 nM, 0.15 μCi). After 10 min, the incubation was stopped with HgCl2 and the mixture was extracted with diethylether. After evaporation, the steroids were dissolved in acetonitrile.

Oster et al.

E1 and E2 were separated using acetonitrile/water (45:55) as mobile phase in a C18 reverse phase chromatography column (Nucleodur C18 Gravity, 3 μm, Macherey-Nagel, D€ uren) connected to a HPLC-system (Agilent 1100 series, Agilent Technologies, Waldbronn). Detection and quantification of the steroids were performed using a radioflow detector (Berthold Technologies, Bad Wildbad). The conversion rate was calculated after analysis of the resulting chromatograms according to the following equation: %E2 %conversion ¼  100 %E2 þ %E1 Each value was calculated from at least three independent experiments. 2. Inhibition of 17β-HSD2. The 17β-HSD2 inhibition assay was performed similarly to the 17β-HSD1 procedure. The microsomal fraction was incubated with NADþ [1500 μM], test compound, and a mixture of unlabeled- and [2,4,6,7-3H]-E2 (final concentration: 500 nM, 0.11 μCi) for 20 min at 37 °C. Further treatment of the samples and HPLC separation was carried out as mentioned above. The conversion rate was calculated after analysis of the resulting chromatograms according to the following equation: %E1 %conversion ¼  100 %E1 þ %E2 3. ER Affinity. The binding affinity of selected compounds to the ERR and ERβ was determined according to Zimmermann et al.54 Briefly, 0.25 pmol of ERR or ERβ, respectively, were incubated with [2,4,6,7-3H]-E2 (10 nM) and test compound for 1 h at room temperature. The potential inhibitors were dissolved in DMSO (5% final concentration). Evaluation of nonspecific-binding was performed with diethylstilbestrol (10 μM). After incubation, ligand-receptor complexes were selectively bound to hydroxyapatite (5 g/60 mL TE-buffer). The complex formed was separated, washed, and resuspended in ethanol. For radiodetection, scintillator cocktail (Quickszint 212, Zinsser Analytic, Frankfurt) was added and samples were measured in a liquid scintillation counter (Rack Beta Primo 1209, Wallac, Turku, Finland). For determination of the relative binding affinity (RBA), inhibitor and E2 concentrations required to displace 50% of the receptor bound labeled E2 were determined. RBA values were calculated according to the following equation: IC50 ðE2Þ  100 RBA½% ¼ IC50 ðcompoundÞ The RBA value for E2 was arbitrarily set at 100%. 4. Inhibition of 17β-HSD1 in T47D Cells. A stock culture of T47D cells was grown in RPMI 1640 medium supplemented with 10% FCS, L-glutamine (2 mM), penicillin (100 IU/mL), streptomycin (100 μg/mL), insulin-zinc-salt (10 μg/mL), and sodium pyruvate (1 mM) at 37 °C under 5% CO2 humidified atmosphere. The cells were seeded into a 24-well plate at 1  106 cells/well in DMEM medium with FCS, and L-glutamine and the antibiotics added in the same concentrations as mentioned above. After 24 h, the medium was changed for fresh serum free DMEM and a solution of test compound in DMSO was added. Final concentration of DMSO was adjusted to 1% in all samples. After a preincubation of 30 min at 37 °C with 5% CO2, the incubation was started by addition of a mixture of unlabeled- and [2,4,6,7-3H]-E1 (final concentration: 50 nM, 0.15 μCi). After 0.5 h incubation, the enzymatic reaction was stopped by removing of the supernatant medium. The steroids were extracted with diethylether. Further treatment of the samples was carried out as mentioned for the 17β-HSD1 assay. 5. Metabolic Stability. Human liver microsomes (final protein concentration 0.5 mg/mL), 0.1 M phosphate buffer pH 7.4, and test compound (final substrate concentration = 3 μM; final

Article

DMSO concentration = 0.25%) were preincubated at 37 °C prior to the addition of NADPH (final concentration=1 mM) to initiate the reaction. Dextromethorphan and verapamil were used as references. All incubations were performed singularly for each test compound. Each compound was incubated for 0, 5, 15, 30, and 45 min (control: 45 min). The reactions were stopped by the addition of 50 μL of methanol containing internal standard at the appropriate time points. The incubation plates were centrifuged at 2000 rpm for 20 min at 4 °C to precipitate the protein. The sample supernatants were combined in cassettes of up to four compounds and analyzed using LC-MS/MS.

Acknowledgment. We are grateful to the Deutsche Forschungsgemeinschaft (HA1315/8-1) for financial support of this work. We thank Jannine Ludwig for her help in performing the in vitro tests. Supporting Information Available: Experimental and spectroscopic data of all compounds as well as HPLC purity determination of all tested compounds. This material is available free of charge via the Internet at http://pubs.acs.org

References (1) Travis, R. C.; Key, T. J. Oestrogen exposure and breast cancer risk. Breast Cancer Res. 2003, 5, 239–247. (2) Dizerega, G. S.; Barber, D. L.; Hodgen, G. D. Endometriosis: role of ovarian steroids in initiation, maintenance and suppression. Fertil. Steril. 1980, 33, 649–653. (3) Herold, C. I.; Blackwell, K. L. Aromatase inhibitors for breast cancer: proven efficacy across the spectrum of disease. Clin. Breast Cancer 2008, 8, 50–64. (4) Gobbi, S.; Zimmer, C.; Belluti, F.; Rampa, A.; Hartmann, R. W.; Recanatini, M.; Bisi, A. Novel highly potent and selective nonsteroidal aromatase inhibitors: synthesis, biological evaluation and structure-activity relationships investigation. J. Med. Chem. 2010, 53, 5347–5351. (5) Le Borgne, M.; Marchand, P.; Delevoye-Seiller, B.; Robert, J. M.; Le Baut, G.; Hartmann, R. W.; Palzer, M. New selective nonsteroidal aromatase inhibitors: synthesis and inhibitory activity of 2,3 or 5-(alpha-azolylbenzyl)-1H-indoles. Bioorg. Med. Chem. Lett. 1999, 9, 333–336. (6) Le Borgne, M.; Marchand, P.; Duflos, M.; Delevoye-Seiller, B.; Piessard-Robert, S.; Le Baut, G.; Hartmann, R. W.; Palzer, M. Synthesis and in vitro evaluation of 3-(1-azolylmethyl)-1H-indoles and 3-(1-azolyl-1-phenylmethyl)-1H-indoles as inhibitors of P450 arom. Arch. Pharm. (Weinheim, Ger.) 1997, 330, 141–145. (7) Schuster, D.; Laggner, C.; Steindl, T. M.; Palusczak, A.; Hartmann, R. W.; Langer, T. Pharmacophore modeling and in silico screening for new P450 19 (aromatase) inhibitors. J. Chem. Inf. Model. 2006, 46, 1301–1311. (8) Leonetti, F.; Favia, A.; Rao, A.; Aliano, R.; Paluszcak, A.; Hartmann, R. W.; Carotti, A. Design, synthesis, and 3D QSAR of novel potent and selective aromatase inhibitors. J. Med. Chem. 2004, 47, 6792–6803. (9) Aidoo-Gyamfi, K.; Cartledge, T.; Shah, K.; Ahmed, S. Estrone sulfatase and its inhibitors. Anticancer Agents Med. Chem. 2009, 9, 599–612. (10) Gunnarsson, C.; Hellqvist, E.; Stal, O. 17beta-Hydroxysteroid dehydrogenases involved in local oestrogen synthesis have prognostic significance in breast cancer. Br. J. Cancer 2005, 92, 547–552. (11) Miyoshi, Y.; Ando, A.; Shiba, E.; Taguchi, T.; Tamaki, Y.; Noguchi, S. Involvement of up-regulation of 17beta-hydroxysteroid dehydrogenase type 1 in maintenance of intratumoral high estradiol levels in postmenopausal breast cancers. Int. J. Cancer 2001, 94, 685–689.   (12) Smuc, T.; Pucelj Ribic, M.; Sinkovec, J.; Husen, B.; Thole, H.; Lanisnik Rizner, T. Expression analysis of the genes involved in estradiol and progesterone action in human ovarian endometriosis. Gynecol. Endocrinol. 2007, 23, 105–111. (13) Aggarwal, S.; Thareja, S.; Verma, A.; Bhardwaj, T. R.; Kumar, M. An overview on 5alpha-reductase inhibitors. Steroids 2010, 75, 109–153. (14) Picard, F.; Schulz, T.; Hartmann, R. W. 5-Phenyl substituted 1-methyl-2-pyridones and 40 -substituted biphenyl-4-carboxylic acids. synthesis and evaluation as inhibitors of steroid-5alphareductase type 1 and 2. Bioorg. Med. Chem. 2002, 10, 437–448.

Journal of Medicinal Chemistry, 2010, Vol. 53, No. 22

8185

(15) Baston, E.; Palusczak, A.; Hartmann, R. W. 6-Substituted 1Hquinolin-2-ones and 2-methoxy-quinolines: synthesis and evaluation as inhibitors of steroid 5alpha reductases types 1 and 2. Eur. J. Med. Chem. 2000, 35, 931–940. (16) Baston, E.; Hartmann, R. W. N-Substituted 4-(5-indolyl)benzoic acids. Synthesis and evaluation of steroid 5alpha-reductase type I and II inhibitory activity. Bioorg. Med. Chem. Lett. 1999, 9, 1601– 1606. (17) Haller, F.; Moman, E.; Hartmann, R. W.; Adamski, J.; Mindnich, R. Molecular framework of steroid/retinoid discrimination in 17beta-hydroxysteroid dehydrogenase type 1 and photoreceptorassociated retinol dehydrogenase. J. Mol. Biol. 2010, 399, 255–267. (18) Marchais-Oberwinkler, S.; Henn, C.; M€ oller, G.; Klein, T.; Lordon, M.; Negri, M.; Oster, A.; Spadaro, A.; Werth, R.; Xu, K.; Frotscher, M.; Hartmann, R. W.; Adamski, J. 17beta-Hydroxysteroid dehydrogenases (17beta-HSD): genes, protein structures, novel therapeutic targets and recent progress in inhibitor development. J. Steroid Biochem. Mol. Biol. 2010, submitted and references therein. (19) Moore, D. M.; Kalvakolanu, D. V.; Lippman, S. M.; Kavanagh, J. J.; Hong, W. K.; Borden, E. C.; Paredes-Espinoza, M.; Krakoff, I. H. Retinoic acid and interferon in human cancer: mechanistic and clinical studies. Semin. Hematol. 1994, 31, 31–37. (20) Husen, B.; Huhtinen, K.; Poutanen, M.; Kangas, L.; Messinger, J.; Thole, H. Evaluation of inhibitors for 17beta-hydroxysteroid dehydrogenase type 1 in vivo in immunodeficient mice inoculated with MCF-7 cells stably expressing the recombinant human enzyme. Mol. Cell. Endocrinol. 2006, 248, 109–113. (21) Husen, B.; Huhtinen, K.; Saloniemi, T.; Messinger, J.; Thole, H. H.; Poutanen, M. Human hydroxysteroid (17-beta) dehydrogenase 1 expression enhances estrogen sensitivity of MCF-7 breast cancer cell xenografts. Endocrinology 2006, 147, 5333–5339. (22) Day, J. M.; Foster, P. A.; Tutill, H. J.; Parsons, M. F.; Newman, S. P.; Chander, S. K.; Allan, G. M.; Lawrence, H. R.; Vicker, N.; Potter, B. V.; Reed, M. J.; Purohit, A. 17beta-Hydroxysteroid dehydrogenase type 1, and not type 12, is a target for endocrine therapy of hormone-dependent breast cancer. Int. J. Cancer 2008, 122, 1931–1940. (23) Lamminen, T.; Saloniemi, T.; Huhtinen, K.; Koskimies, P.; Messinger, J.; Husen, B.; Thole, H.; Poutanen, M. In vivo mouse model for analysis of hydroxysteroid (17beta) dehydrogenase 1 inhibitors. Mol. Cell. Endocrinol. 2009, 301, 158–162. (24) Gr€ ummer, R.; Schwarzer, F.; Bainczyk, K.; Hess-Stumpp, H.; Regidor, P. A.; Schindler, A. E.; Winterhager, E. Peritoneal endometriosis: validation of an in vivo model. Hum. Reprod. 2001, 16, 1736–1743. (25) Einspanier, A.; Lieder, K.; Bruns, A.; Husen, B.; Thole, H.; Simon, C. Induction of endometriosis in the marmoset monkey (Callithrix jacchus). Mol. Hum. Reprod. 2006, 12, 291–299. (26) Poirier, D. Advances in development of inhibitors of 17beta hydroxysteroid dehydrogenases. Anticancer Agents Med. Chem. 2009, 9, 642–660 and references therein. (27) Brozic, P.; Lanisnik Rizner, T.; Gobec, S. Inhibitors of 17betahydroxysteroid dehydrogenase type 1. Curr. Med. Chem. 2008, 15, 137–150 and references therein. (28) Day, J. M.; Tutill, H. J.; Purohit, A.; Reed, M. J. Design and validation of specific inhibitors of 17beta-hydroxysteroid dehydrogenases for therapeutic application in breast and prostate cancer, and in endometriosis. Endocr. Relat. Cancer 2008, 15, 665–692 and references therein. (29) Schuster, D.; Nashev, L. G.; Kirchmair, J.; Laggner, C.; Wolber, G.; Langer, T.; Odermatt, A. Discovery of nonsteroidal 17betahydroxysteroid dehydrogenase 1 inhibitors by pharmacophorebased screening of virtual compound libraries. J. Med. Chem. 2008, 51, 4188–4199. (30) Michiels, P. J. A.; Ludwig, C.; Stephan, M.; Fischer, C.; M€ oller, G.; Adamski, J.; van Dongen, M.; Thole, H.; G€ unther, U. L. Ligand based NMR spectra demonstrate an additional phytoestrogen binding site for 17 beta-hydroxysteroid dehydrogenase type-1. J. Steroid Biochem. Mol. Biol. 2009, 117, 93–98. (31) Negri, M.; Recanatini, M.; Hartmann, R. W. Insights in 17betaHSD1 enzyme kinetics and ligand binding by dynamic motion investigation. PLoS One 2010, 5, e 12026. (32) Brozic, P.; Kocbek, P.; Sova, M.; Kristl, J.; Martens, S.; Adamski, J.; Gobec, S.; Lanisnik Rizner, T. Flavonoids and cinnamic acid derivatives as inhibitors of 17beta-hydroxysteroid dehydrogenase type 1. Mol. Cell. Endocrinol. 2009, 301, 229–234. (33) Messinger, J.; Hirvel€a, L.; Husen, B.; Kangas, L.; Koskimies, P.; Pentik€ainen, O.; Saarenketo, P.; Thole, H. New inhibitors of 17betahydroxysteroid dehydrogenase type 1. Mol. Cell. Endocrinol. 2006, 248, 192–198.

8186 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 22 (34) Karkola, S.; Lilienkampf, A.; W€ah€al€a, K. A 3D QSAR model of 17beta-HSD1 inhibitors based on a thieno[2,3-d]pyrimidin-4(3H)one core applying molecular dynamics simulations and ligandprotein docking. ChemMedChem 2008, 3, 461–472. (35) Frotscher, M.; Ziegler, E.; Marchais-Oberwinkler, S.; Kruchten, P.; Neugebauer, A.; Fetzer, L.; Scherer, C.; M€ uller-Vieira, U.; Messinger, J.; Thole, H.; Hartmann, R. W. Design, synthesis and biological evaluation of (hydroxyphenyl)-naphthalene and quinoline derivatives: potent and selective nonsteroidal inhibitors of 17beta-hydroxysteroid dehydrogenase type 1 (17beta-HSD1) for the treatment of estrogen-dependent diseases. J. Med. Chem. 2008, 51, 2158–2169. (36) Marchais-Oberwinkler, S.; Kruchten, P.; Frotscher, M.; Ziegler, E.; Neugebauer, A.; Bhoga, U. D.; Bey, E.; M€ uller-Vieira, U.; Messinger, J.; Thole, H.; Hartmann, R. W. Substituted 6-phenyl-2naphthols. Potent and selective nonsteroidal inhibitors of 17βhydroxysteroid dehydrogenase type 1 (17beta-HSD1): design, synthesis, biological evaluation and pharmacokinetics. J. Med. Chem. 2008, 51, 4685–4698. (37) Marchais-Oberwinkler, S.; Frotscher, M.; Ziegler, E.; Werth, R.; Kruchten, P.; Messinger, J.; Thole, H.; Hartmann, R. W. Structureactivity study in the class of 6-(30 -hydroxyphenyl)naphthalenes leading to an optimization of a pharmacophore model for 17betahydroxysteroid dehydrogenase type 1 (17beta-HSD1) inhibitors. Mol. Cell. Endocrinol. 2009, 301, 205–211. (38) Kruchten, P.; Werth, R.; Marchais-Oberwinkler, S.; Frotscher, M.; Hartmann, R. W. Development of a biological screening system for the evaluation of highly active and selective 17beta-HSD1-inhibitors as potential therapeutic agents. Mol. Cell. Endocrinol. 2009, 301, 154–157. (39) Marchais-Oberwinkler, S.; Wetzel, M.; Ziegler, E.; Kruchten, P.; Werth, R.; Hartmann, R. W.; Frotscher, M. New drug-like hydroxyphenylnaphthol steroidomimetics as potent and selective 17betaHSD1 inhibitors for the treatment of estrogen dependent diseases. J. Med. Chem. 2010, submitted for publication. (40) Bey, E.; Marchais-Oberwinkler, S.; Kruchten, P.; Frotscher, M.; Werth, R.; Oster, A.; Algul, O.; Neugebauer, A.; Hartmann, R. W. Design, synthesis and biological evaluation of bis(hydroxyphenyl) azoles as potent and selective nonsteroidal inhibitors of 17betahydroxysteroid dehydrogenase type 1 (17beta-HSD1) for the treatment of estrogen-dependent diseases. Bioorg. Med. Chem. 2008, 16, 6423–6435. (41) Bey, E.; Marchais-Oberwinkler, S.; Werth, R.; Negri, M.; Al-Soud, Y. A.; Kruchten, P.; Oster, A.; Frotscher, M.; Birk, B.; Hartmann, R. W. Design, synthesis, biological evaluation and pharmacokinetics of bis(hydroxyphenyl) substituted azoles, thiophenes, benzenes and aza-benzenes as potent and selective non-steroidal inhibitors of 17betahydroxysteroid dehydrogenase type 1 (17beta-HSD1). J. Med. Chem. 2008, 51, 6725–6739. (42) Al-Soud, Y. A.; Bey, E.; Oster, A.; Marchais-Oberwinkler, S.; Werth, R.; Kruchten, P.; Frotscher, M.; Hartmann, R. W. The role of the heterocycle in bis(hydroxyphenyl)triazoles for inhibition of

Oster et al.

(43)

(44)

(45)

(46)

(47)

(48) (49)

(50)

(51) (52)

(53)

(54)

17beta-hydroxysteroid dehydrogenase (17beta-HSD) type 1 and type 2. Mol. Cell. Endocrinol. 2009, 301, 212–215. Bey, E.; Marchais-Oberwinkler, S.; Negri, M.; Kruchten, P.; Oster, A.; Werth, R.; Frotscher, M.; Birk, B.; Hartmann, R. W. New insights into the SAR and binding modes of bis(hydroxyphenyl)thiophenes and benzenes: influence of additional substituents on 17β-hydroxysteroid dehydrogenase type 1 (17β-HSD1) inhibitory activity and selectivity. J. Med. Chem. 2009, 52, 6724–6743. Kruchten, P.; Werth, R.; Bey, E.; Oster, A.; Marchais-Oberwinkler, S.; Frotscher, M.; Hartmann, R. W. Selective inhibition of 17betahydroxysteroid dehydrogenase type 1 (17betaHSD1) reduces estrogen responsive cell growth of T47-D breast cancer cells. J. Steroid Biochem. Mol. Biol. 2009, 114, 200–206. Kruchten, P.; Werth, R.; Marchais-Oberwinkler, S.; Bey, E.; Ziegler, E.; Oster, A.; Frotscher, M.; Hartmann, R. W. Development of biological assays for the identification of selective inhibitors of estradiol formation from estrone in rat liver preparations. C. R. Chim. 2009, 12, 1110–1116. Oster, A.; Klein, T.; Werth, R.; Kruchten, P.; Bey, E.; Negri, M.; Marchais-Oberwinkler, S.; Frotscher, M.; Hartmann, R. W. Novel estrone mimetics with high 17beta-HSD1 inhibitory activity. Bioorg. Med. Chem. 2010, 18, 3494–3505. Mazumdar, M.; Fournier, D.; Zhu, D. W.; Cadot, C.; Poirier, D.; Lin, S. X. Binary and ternary crystal structure analyses of a novel inhibitor with 17beta-HSD type 1: a lead compound for breast cancer therapy. Biochem. J. 2009, 424, 357–366. Miyaura, N.; Suzuki, A. The palladium catalysed cross-coupling reaction of phenylboronic acid with haloarenes in the presence of bases. Synth. Commun. 1995, 11, 513–519. Fink, B. E.; Mortensen, D. S.; Stauffer, S. R.; Aron, Z. D.; Katzenellenbogen, J. A. Novel structural templates for estrogenreceptor ligands and prospects for combinatorial synthesis of estrogens. Chem. Biol. 1999, 6, 205–219. Allan, G. M.; Vicker, N.; Lawrence, H. R.; Tutill, H. J.; Day, J. M.; Huchet, M.; Ferrandis, E.; Reed, M. J.; Purohit, A.; Potter, B. V. Novel inhibitors of 17beta-hydroxysteroid dehydrogenase type 1: templates for design. Bioorg. Med. Chem. 2008, 16, 4438–4456. Bhatt, M. V.; Kulkarni, S. U. Cleavage of ethers. Synthesis 1983, 249–282. Czaplewski, L. G.; Collins, I.; Boyd, E. A.; Brown, D.; East, S. P.; Gardiner, M.; Fletcher, R.; Haydon, D. J.; Henstock, V.; Ingram, P.; Jones, C.; Noula, C.; Kennison, L.; Rockley, C.; Rose, V.; ThomaidesBrears, H. B.; Ure, R.; Whittaker, M.; Stokes, N. R. Antibacterial alkoxybenzamide inhibitors of the essential bacterial cell division protein FtsZ. Bioorg. Med. Chem. Lett. 2009, 19, 524–527. Morgan, B. P.; Galdamez, G. A.; Gilliard, R. J., Jr.; Smith, R. C. Canopied trans-chelating bis(N-heterocyclic carbene) ligand: synthesis, structure and catalysis. Dalton Trans. 2009, 2020– 2028. Zimmermann, J.; Liebl, R.; von Angerer, E. 2,5-Diphenylfuranbased pure antiestrogens with selectivity for the estrogen receptor alpha. J. Steroid Biochem. Mol. Biol. 2005, 94, 57–66.